You get a frantic phone call at 3am on a Sunday. It’s your boss, and there’s been a fire. The server room, and perhaps the whole building, is still smoldering. He asks you the following questions:
- Is our data safe?
- Is it off site? Is it secure or on someone’s kitchen table?
- When was the last time you checked the back up? You are checking them, right?
- Are you sure you can restore the M2M files? You have been performing test restores, right?
There are many blog articles on the web about the importance of disaster recovery, and you’ve probably read a few of them. I won’t re-hash everything here, but do you have the answers to all of those questions? If not (and you value your job), stop reading this and address them immediately. I’ll wait….
Good! Now that the above are covered lets address some M2M specific issues.
- Can you build your SQL Server and install everything without help? Servers seem to prefer to die on weekends, and M2M Support is not available late at night or on weekends and holidays.
- Where is your software? Latest Service Packs? Software keys? At the bare minimum, a M2M server requires Windows Server, Visual FoxPro, SQL, and Made2Manage.
- Do you have documentation handy with all of your M2M and other settings?
- Do you have instructions ready to install all of your optional modules? Do you have the latest install files for them? I recommend keeping all of this (including the install files for everything) in electronic format with your back-ups if possible.
I remember watching a movie where special forces soldiers spent hours assembling and disassembling their rifles blindfolded. The point was that they became intimately familiar with their weapon and knew everything about it. Ultimately, this made them more effective soldiers.
Recently I wrote disaster recovery instructions for SQL Server, Made2Manage, and the associated modules for my current employer. One of my fellow employees performed the same process on a Domain Controller and Exchange Server. We then verified them by restoring everything from bare hardware, back-ups, and our instructions in an isolated room without access to the internet. Our manager also put us under stress by telling us our results were going to be public and that we had better not let him down.
Why a closed room and stress? It simulates a disaster. Can you be assured that you will have access to the internet in a disaster? Even if you could, would your boss want to wait while you download service packs? Since we had all of the files and had practiced, we built and tested our servers in less than 4 hours.
What are the benefits of a complete test?
- To prove we could do it.
- I learned exactly what makes M2M work.
- I’m now confident that I can fix anything on that server.
Incidentally, with the exception of our database server, the rest of our servers are replicated in two other sites in the United States. In the near future, our SQL Server will be replicated the same way. So we should never actually be “under the gun” so to speak.
So Hotshot, what do you do?